Paste a URL and we'll run three independent AI-readiness checks in parallel: AI crawler access, llms.txt validity, and JSON-LD schema. See where you stand and what to fix.
AI systems can only cite what they can crawl, interpret, and trust. These three checks test the foundations.
1
Access
AI crawler access
We check whether major AI crawlers are allowed to access your site.
Why it matters
If they can't access, they can't discover your content.
2
Instructions
llms.txt validity
We verify your machine-readable instructions file is present and valid.
Why it matters
Clear instructions help AI systems use your content correctly.
3
Understanding
JSON-LD schema for AI
We evaluate structured data so AI systems can understand your pages.
Why it matters
Structured data improves context, attribution and visibility.
Run the diagnostics
Reads the target site's /robots.txt, parses the User-agent + Allow/Disallow blocks, then cross-checks meta robots tags and the X-Robots-Tag response header. Reports per-bot access for the 19 named crawlers we track: GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-Web, anthropic-ai, PerplexityBot, Perplexity-User, xAI-Crawler, Google-Extended, Googlebot, Applebot, Applebot-Extended, Bingbot, CCBot, Bytespider, DuckAssistBot, Diffbot, Cohere-AI. A bot is reported blocked if any of those three layers excludes it. If your scan reports any crawler as blocked, the cause is usually a wildcard rule that pre-dates generative search.
llms.txt is a markdown file at the root of a domain that tells AI assistants which pages on the site they should read first. Format proposed by Jeremy Howard in 2024, adopted by Anthropic, Cloudflare, Vercel, and thousands of other sites. Structure: H1 site name, blockquote description, sections grouping canonical pages by topic. The validator checks both /llms.txt at apex and /.well-known/llms.txt, parses the structure, HEAD-checks a 15-link sample for integrity, and grades A / B / C on presence + format + link health.
JSON-LD is the structured-data format Google, Bing, and AI engines parse before reading any prose on a page. Seven types drive AI citation: Organization (entity recognition), Person (E-E-A-T author signal), FAQPage (direct Q&A pulls), HowTo (step-by-step citations), Product (shopping queries), Article (content attribution), BreadcrumbList (site structure). The checker fetches the URL, parses every application/ld+json block, and reports per-type field coverage against the AI-weighted floor. Common gaps: missing sameAs on Organization, missing jobTitle + worksFor on Person, FAQPage answers under 50 words.
What the three checks answer
Three plain-English questions.
If any answer is no, the AI doesn't see you. If all three are yes, you're cleared for citation.
Access
Can AI systems access your site?
We confirm AI crawlers are allowed through robots.txt and aren't blocked by meta rules.
Instructions
Can they find your instructions?
We verify your llms.txt is present and valid, so AI systems know how to use your content.
Understanding
Can they understand your pages?
We assess your JSON-LD schema so AI systems can understand meaning, entities and context.
How this tool fits with Pulse
This checks the foundation. Pulse checks the full picture.
The free tool shows whether AI systems can access and understand your site. Pulse scores the wider visibility system: authority, content quality, structured data, platform presence, AI citability, and readiness to be recommended.
Same checks as a Pulse audit
These three probes are deterministic checks lifted directly from the Pulse audit pipeline. The tools surface what we score; the full audit interprets and prioritises across seven categories.
No upsell, no email
Run as many scans as you want. We log nothing per scan. There is no signup wall, no email gate, no contact form on submit.
When to commission an audit
These tools cover three of the seven Pulse audit categories. The other four (Brand Authority, Content Quality, Technical Foundations, Platform Optimisation) need a full audit.
